提交 · 50cafa0b0c03116903016552630a818230cce003 · Crayon鑫 / Paddle

19 3月, 2021 2 次提交
- Z
  
  remove redundant sync, set collect/dist kernel to context stream, sub_lod memcpy opt (#31641) · 50cafa0b
  由 zlsh80826 提交于 3月 19, 2021
  
  50cafa0b
- R
  
  [ROCM] fix layer_norm, norm, p_norm, test_sequence_softmax_op, test_math_op_patch_var_base (#31709) · 420527f0
  由 ronnywang 提交于 3月 19, 2021
  
  420527f0
17 3月, 2021 1 次提交
- Z
  
  support NHWC for temporal_shift op (#31642) · 7f50bb7e
  由 Zhang Ting 提交于 3月 17, 2021
  
  7f50bb7e
16 3月, 2021 1 次提交
- R
  
  [ROCM] fix softmax_with_cross_entropy_op, test=develop (#31629) · da10c5cf
  由 ronnywang 提交于 3月 16, 2021
  
  da10c5cf
15 3月, 2021 2 次提交
- W
  
  c_gen_nccl_id add SocketServer to persit server (#31589) · 9066b74f
  由 WangXi 提交于 3月 15, 2021
  
  9066b74f
- K
  DataLoader supprot dict str (#31481) · a32e8bf1
  由 Kaipeng Deng 提交于 3月 15, 2021
```
* add dict/str/list supprot for DataLoader. test=develop
```
  a32e8bf1
12 3月, 2021 2 次提交
- P
  [Paddle-TRT] Fix engine key in trt int8 calibration (#31513) · cac9635a
  由 Pei Yang 提交于 3月 12, 2021
```
* fix engine key in trt int8 calibration

* fix unit test
```
  cac9635a
- Q
  
  [ROCM] fix conv2d and conv3d op, test=develop (#31553) · 3d5aa9d1
  由 Qi Li 提交于 3月 12, 2021
  
  3d5aa9d1
11 3月, 2021 2 次提交
- J
  
  optimize range op by place parameters on cpu rather than gpu, test=develop (#30811) · 9ed6c895
  由 jiangcheng 提交于 3月 11, 2021
  
  9ed6c895
- C
  add softmax_switch for softmax_with_cross_entropy_op, test=develop (#31428) · 6148b87f
  由 chajchaj 提交于 3月 11, 2021
```
* add softmax_switch for softmax_with_cross_entropy_op, test=develop

* delete using EigenMatrix in softmax_with_cross_entropy_op.h, test=develop

* add REGISTER_OP_VERSION for softmax_switch attr of softmax_with_cross_entropy_op, test=develop
```
  6148b87f
10 3月, 2021 5 次提交
- W
  
  Add collective async wait op (#31463) · 83a2fb1f
  由 WangXi 提交于 3月 10, 2021
  
  83a2fb1f
- L
  remove the send/recv of tensor size (#31460) · 0205e9f8
  由 lilong12 提交于 3月 10, 2021
```
* remove the send/recv of tensor size, but users have to specify the shape of the received var explicitly.
```
  0205e9f8
- F
  Bugfix rocm (#31490) · 910f377f
  由 furnace 提交于 3月 10, 2021
```
* bugfix for test_cholesky_op

* bugfix for test_compare_op

* bugfix for lookup_table_op

* bugfix for affine_channel_op
```
  910f377f
- Q
  
  [ROCM] fix softmax with loss nan in HIP platform, test=develop (#31491) · 416e47ed
  由 Qi Li 提交于 3月 10, 2021
  
  416e47ed
- J
  
  Optimization of elementwise CUDA kernel (#30801) · 45c7d905
  由 JamesLim 提交于 3月 10, 2021
  
  45c7d905
09 3月, 2021 6 次提交
- J
  [oneDNN] bumpup onednn 2.2 fixup version (#31473) · 23d96cf2
  由 Jacek Czaja 提交于 3月 09, 2021
```
* - introduced fix onednn 2.2 version

* - compilation fix
```
  23d96cf2
- W
  
  update conv2d, test=develop (#31480) · 43d6abf0
  由 wangguanzhong 提交于 3月 09, 2021
  
  43d6abf0
- W
  
  fix roi_align, test=develop (#31479) · 50af0c2c
  由 wangguanzhong 提交于 3月 09, 2021
  
  50af0c2c
- R
  
  [ROCM] fix gather_op, sigmoid_cross_entropy_with_logits_op, test=develop (#31467) · e03e4673
  由 ronnywang 提交于 3月 09, 2021
  
  e03e4673
- Q
  
  [ROCM] fix reduce op, test=develop (#31478) · b85c8e03
  由 Qi Li 提交于 3月 09, 2021
  
  b85c8e03
- J
  
  [oneDNN] elementwise add bf16 grad kernel with broadcasting (#31385) · 39a5424e
  由 Jacek Czaja 提交于 3月 09, 2021
  
  39a5424e
08 3月, 2021 2 次提交
- Q
  
  [ROCM] fix test_dist_op ci test, test=develop (#31468) · 133a914b
  由 Qi Li 提交于 3月 08, 2021
  
  133a914b
- Q
  
  [ROCM] fix dropout and remove hipcub, test=develop (#31455) · f9377965
  由 Qi Li 提交于 3月 08, 2021
  
  f9377965
05 3月, 2021 2 次提交
- J
  
  Creating a CUDA function to find the minimum value in warp or block (#31191) · 8491ae9a
  由 JamesLim 提交于 3月 05, 2021
  
  8491ae9a
- L
  [Kunlun]Multi xpu dygraph performance optimization , add distributed.spawn... · 9ebf05b0
  由 liuyuhui 提交于 3月 05, 2021
```
[Kunlun]Multi xpu dygraph performance optimization , add distributed.spawn support for multi xpu and some bug-fixes (#31130)
```
  9ebf05b0
04 3月, 2021 4 次提交
- Q
  
  [ROCM] update fluid platform for rocm (part5), test=develop (#31315) · 4d647ec1
  由 Qi Li 提交于 3月 04, 2021
  
  4d647ec1
- Z
  
  support float16 for temporal_shift op (#31432) · 7d95e598
  由 Zhang Ting 提交于 3月 04, 2021
  
  7d95e598
- L
  
  Fix bug for set_value op when input dtype is not float32 (#31411) · 0fff9306
  由 liym27 提交于 3月 04, 2021
  
  0fff9306
- J
  
  Added LSTM BF16 and fixed GRU BF16 (#31234) · 5b4f8aac
  由 jakpiase 提交于 3月 04, 2021
  
  5b4f8aac
03 3月, 2021 7 次提交
- Q
  [ROCM] update fluid elementwise op for rocm (part10), test=develop (#31361) · 7cdf6ea7
  由 Qi Li 提交于 3月 03, 2021
```
* [ROCM] update fluid elementwise op for rocm (part10), test=develop

* update, test=develop

* address review comments, test=develop
```
  7cdf6ea7
- Q
  [ROCM] update fluid operators for rocm (part3), test=develop (#31213) · 84639b61
  由 Qi Li 提交于 3月 03, 2021
```
* [ROCM] update fluid operators for rocm (part3), test=develop

* fix clang format error, test=develop
```
  84639b61
- Q
  
  [ROCM] update fluid operators for rocm (part7), test=develop (#31307) · 3b9db171
  由 Qi Li 提交于 3月 03, 2021
  
  3b9db171
- Q
  
  [ROCM] fix softmax with loss and update python scripts, test=develop (#31373) · db50fb67
  由 Qi Li 提交于 3月 03, 2021
  
  db50fb67
- Q
  
  [ROCM] update fluid operators for rocm (part9), test=develop (#31338) · e312a1ff
  由 Qi Li 提交于 3月 03, 2021
  
  e312a1ff
- Q
  
  fix bert cu file compiler error, test=develop (#31389) · 6626c6a6
  由 Qi Li 提交于 3月 03, 2021
  
  6626c6a6
- Q
  
  [ROCM] update fluid operators for rocm (part6), test=develop (#31301) · 946dbdae
  由 Qi Li 提交于 3月 03, 2021
  
  946dbdae
02 3月, 2021 4 次提交

Q

[ROCM] update fluid operators for rocm (part8), test=develop (#31309) · 59940cb3
由 Qi Li 提交于 3月 02, 2021

59940cb3
Q

fix ELU output for nan, test=develop (#31132) · ec72f5b2
由 Qi Li 提交于 3月 02, 2021

ec72f5b2

[ROCM] update fluid operators for rocm (part5), test=develop (#31258) · 65bcaeb0

由 Qi Li 提交于 3月 02, 2021

* [ROCM] update fluid operators for rocm (part5), test=develop

* address review comments, test=develop

* fix typo, test=develop

65bcaeb0

lamb_op_xpu;test=kunlun (#31012) · d79fdc3d

由 Gradie 提交于 3月 02, 2021

* lamb_op_xpu;test=kunlun

* modify lamb_op_xpu.cc;test=kunlun

* delete atol lamb_op_xpu; test=kunlun

* update xpu.cmake;test=kunlun

* test_error 1e-5,lamb_op_xpu;test=kunlun

* error1e-5,lamb_op_xpu,test=kunlun

* delete atol lamb_xpu;test=kunlun

* modify atol,lamb_op_xpy;test=kunlun

* lamb_op_xpu;test=kunlun

* lamb_op_xpu;test=kunlun

* lamb_op_xpu, XPUOptest;test=kunlun

* lamb_op_xpu;test=kunlun

* lamb_op_xpu;test=kunlun

* lamb_op_xpu;test=kunlun

* lamb_op_xpu;test=kunlun

* lamb_op_xpu;test=kunlun

* lamb_op_xpu;test=kunlun

* lamb_op_xpu;test=kunlun

* lamb_op_xpu;test=kunlun

* lamb_op_xpu;test=kunlun

* lamb_op_xpu;test=kunlun

* lamb_op_xpu;test=kunlun

* lamb_op_xpu,modify xpu_cmake; test=kunlun

* lamb_op_xpu;test=kunlun

* lamb_op_xpu,modify xpucmake;test=kunlun

d79fdc3d

Crayon鑫 / Paddle 与 Fork 源项目一致

Crayon鑫 / Paddle
与 Fork 源项目一致